Work in Progress

Perceptual systems regulate our beliefs in response to experience by generating signals that track features of the environment. I ask what, if anything, plays an analogous role for desire. Drawing on computational cognitive science, I argue that we possess a quasi-perceptual capacity that I call conaception. This capacity functions as a feature-detector, generating non-conceptual value signals that update subpersonal value representations governing our desires, much as perceptual signals update belief. I further argue that conaception is a reliable process: it tracks what is good for us—what is conducive to our well-being—in much the same way perceptual capacities reliably track properties such as distance. Because conaception is typically accurate, desires shaped by it enjoy a defeasible normative authority, just as beliefs generated by our perceptual capacities do. By contrast, desires produced by motivational illusions—cases in which conaception misfires—or by other unreliable processes, such as some mental illnesses, lack this normative authority. This explains why we are usually, but not always, rational in relying on our ordinary desires, and how engaging in practical deliberation can allow us to improve them.

There is a common assumption that humans and other agents learn what to do from the valence of their experiences. This view has seemingly been vindicated by recent developments in computational cognitive science, apparently implying motivational hedonism. In this paper, I argue that this view is based on a conflation between two distinct notions of valence: learning valence, which determines future attraction or aversion, and hedonic valence, the felt quality of pleasure or displeasure. I present a range of evidence that these refer to different mental kinds. This means that, contrary to the suggestion above, motivational hedonism is false. It also means that the capacity to learn from trial and error is no evidence of sentience, contrary to some arguments for attributing sentience to simple animals.

At the heart of my dissertation is a new view of motivation that I call externalism about expected value maximization, or externalism for short. Externalism agrees with standard accounts of rationality that a rational agent performs actions that maximize the achievement of some metric of value. However, it relaxes an additional assumption that is typically made implicitly, namely that what has value depends on the agent themselves, for example, their preferences or desires. According to externalism, rational agents represent value as they would other quantities: as existing independent of themselves and to be estimated in light of evidence. Unlike estimations of other quantities, estimations of value govern their behavior: motivation is the direct pursuit of expected value. 'Rationality Externalized' is the introductory essay of the dissertation, presenting this idea.